Re: Decision time.... should NATA become the default for this release?

[Sascha Wildner <saw@xxxxxxxxx>]

1D31E.6030805@online.de> <200706022051.l52Kp9IY038754@apollo.backplane.com>
In-Reply-To: <200706022051.l52Kp9IY038754@apollo.backplane.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 52
NNTP-Posting-Host: 87.78.98.243
X-Trace: 1180821397 crater_reader.dragonflybsd.org 794 87.78.98.243
Xref: crater_reader.dragonflybsd.org dragonfly.kernel:11102

Matthew Dillon wrote:
> :Nope, sorry. Still the same result (rcorder segfaults etc.).
> :
> :Sascha
> :
> :-- 
> :http://yoyodyne.ath.cx
> 
>     It's got to be reg48, reg4a, or reg54.
> 
>     Try this patch.  Note that its relative /usr/src/sys/dev/disk.  I added
>     printf's to both ATA and NATA, please tell me what the output says for
>     both the NATA and ATA kernels.
> 
>     I also changed another pci_write_config, so it might magically work
>     (tell me if it does or doesn't).

The patch contains the ata-chipset.c part twice (as nata/ata-chipset.c 
and as ata-chipset.c). After removing the last one, it applied cleanly 
from sys/dev/disk.

But it still doesn't work. :(

NATA output:

. ..
ad0: setting PIO4 on PIIX4 chip
ad0: regs before 40=e303e307 44=00 48=05 4a=0202 54=0000
ad0: regs after 40=e377e377 44=00 48=04 4a=0200 54=0000
ad0: setting UDMA33 on PIIX4 chip
ad0: regs before 40=e377e377 44=00 48=04 4a=0200 54=0000
ad0: regs after 40=e377e377 44=00 48=05 4a=0201 54=0000
. ..
acd0: setting PIO4 on PIIX4 chip
acd0: regs before 40=e377e377 44=00 48=05 4a=0201 54=0000
acd0: regs after 40=e377e377 44=00 48=01 4a=0001 54=0000
acd0: setting UDMA33 on PIIX4 chip
acd0: regs before 40=e377e377 44=00 48=01 4a=0001 54=0000
acd0: regs after 40=e377e377 44=00 48=05 4a=0101 54=0000
. ..

ATA output:

. ..
ad0: success setting UDMA2 on Intel chip
ata0: regs after 40=e303e307 44=00 48=05 4a=0202 54=0000
. ..

Sascha

-- 
http://yoyodyne.ath.cx